February 08, 2009, 05:45 PM ET

More Evidence Against Grades

Help! My ‘A’ is melting!

One of the periodic embarrassments for the public schools comes every few years when state tests are administered, scores are tabulated, and officials trumpet the good news.

“The average went up!” they proclaim.

“XX percent of our kids are proficient in math, and YY percent in reading.”

“Program Z is working!”

But then come the NAEP scores for the state, and the gains go down, and sometimes disappear.

A couple of years ago, Diane Ravitch summarized the problem:

“Almost all states report that, based on their own tests, incredibly large proportions of their students meet high standards. Yet the scores on the federal test . . . were far lower. Basically, the states have embraced low standards and grade inflation.”

Some of the discrepancies Ravitch cited were ludicrous.

“Idaho claims that 90 percent of its fourth-grade students are proficient in mathematics, but on the federal test only 41 percent reached the Education Department’s standard of proficiency. Similarly, New York reports that nearly 85 percent of its fourth graders meet state standards in mathematics, yet only 36 percent tested as proficient on the national assessment. North Carolina boasts an impressive 92 percent pass rate on the state test, but only 40 percent meet the federal standard.”

Ravitch and others have called for more uniformity in scales among the states, but a report that came out today in my local paper, The Atlanta Journal-Constitution, suggests that the problem of exaggeration goes a lot deeper than state test scores. In fact, state assessments in Georgia look downright rigorous compared to another element in the evaluation network: classroom grades.

The headline: “Marks from teachers, test scores vary widely: In many cases, ‘A’ student can get ‘F’ on state’s exam.”

Some new discrepancies: While only 6 percent of students in the state failed an economics class, fully 36 percent of students failed the test. In one Atlanta high school, while one-quarter of the students failed geometry, 85 percent of them failed the End of Course Test. Overall, the portion of students who failed most of the tests exceeded the portion of those who failed the classes by two to three times.

Another example of assigned grades reaching the point of meaninglessness, and more evidence for a different model of ranking and rating students. We need tests as a check on grade distortion, but more narrative evaluation, too.